1. Field of the Invention
The present invention relates to a communication path designing method and a communication path designing device to design a communication path to be used according to a service class on a network having a plurality of links that can receive packets each being handled according to a different service class and a plurality of nodes that can differentiate packets according to a service class and can receive the differentiated packets.
The present application claims priority of Japanese Patent Application No.2001-165036 filed on May 31, 2001, which is hereby incorporated by reference.
2. Description of the Related Art
As a mechanism to control QoS (Quality of Service) in throughput, delay in transmission of data, reliability in communications or a like for end-to-end communications, for example, “DiffServ” (Differential Services) technology being now under a process of standardization by IETF (Internet Engineering Task Force) is known.
The DiffServ is used to differentiate packets that flow in and out from a network. In the DiffServ technology, each packet is provided with a specified “Codepoint” corresponding to a service class (precedence) which is assigned based on an SLA (Service Level Agreement) agreed between a manager of a network and a user of the network for transmission of packets.
The Codepoint employed in the DiffServ technology is used to identify precedence to be given to a packet and is made up of a bit pattern indicating a service class including the class to be processed by an EF (Expedited Forwarding) method, AF (Assured Forwarding) methods (including the methods 1–4), or BF (Best Effort Forwarding) method. The EF method class can provide highest precedence, which is used for forwarding of voice data or for transfer data using a virtual IP (Internet Protocol) private line. The BF method class provides lowest precedence, which is used in a Best—Effort—type data transfer in which the QoS during transmission is not guaranteed.
The AF method class provides four types of precedence (AF1, AF2, AF3, and AF4) which are classified as intermediate precedence between the EF method and BF method classes. For example, the AF1 has highest precedence and the precedence decreases in order of AF2 to AF4.
As protocols for receiving and transmitting data in the Internet, a TCP (Transmission Control Protocol) and a UDP (User Datagram Protocol) are known. In the case of the TCP, after a packet is transmitted, an Acknowledge message from a destination to which the packet was transmitted is waited for and then a subsequent packet is forwarded. In the case of the UDP, after a packet is transmitted, a packet is continuously forwarded without waiting for an Acknowledge message from a destination. The service class described above also can be assigned to each of the TCP and UDP.
When a packet or a like is transmitted between an arbitrary node and another node both making up a network, a communication path is designed in a manner so as to meet a specified traffic demand. The traffic demand represents a requirement for a communication path to be created on a network to transmit a packet or a like in order to achieve forwarding of a packet or a like in accordance with contents agreed in the above SLA including, for example, a transfer band, delay time, costs or a like. The communication path designing device selects a communication path that can meet the traffic demand.
Conventionally, when such the communication path as described above on a network based on specified requirements is designed, a designing method disclosed, for example, in Japanese Patent Application Laid-open No. Hei 6-90235 or No. Hei 9-83546 has been used.
However, in the designing method disclosed in the above applications, no specified service class as described above is designated. Therefore, in order to design a communication path that can satisfy contents specified as a service class, the designing has to be made in a manner that most rigorous conditions required to satisfy contents specified as a service class are to be met or in a manner that links or nodes being components making up the network are assigned to satisfy each of contents designated as the service class.
However, the above conventional method for designing a communication path in which the designing has to be made in a manner that most rigorous conditions required to satisfy contents specified as a service class are to be met has a problem in that excessive quality of service is provided to satisfy the traffic demand for a part of service classes, which, as a result, causes wasteful consumption of a resource of a network.
Also, the above conventional method for designing a communication path in which the designing has to be made in a manner that links or nodes being components making up the network are assigned to satisfy each of contents designated as a service class has a problem in that, even if resources assigned to satisfy contents designated as the service class are sufficiently available, since the assigned resources cannot be used, as a substitutional one, to satisfy contents specified according to other service class, which causes impediment to effective utilization of network resources.